engblogs

Discover how to identify text generated by large language models using the Ghostbuster technique.

Join the ElevenLabs Creator Affiliate Program and earn a 22% commission on every payment your referrals make in their first year.

Improved DDIM sampling using Moment Matching Gaussian Mixtures

Improving vision-inspired keyword spotting using a streaming Conformer encoder with input-dependent dynamic depth.

PLANNER: Generating diversified and fluent paragraphs with latent language diffusion model

Future-Driven Dev Empowerment: Slack SVP Tackles New AI and Automation Challenges Head-On

Discover the new features in Mojo SDK v0.5, including keyword parameters, automatic parameterization of functions, tensor and string enhancements, and benchmark enhancements.

Scaling multimodal understanding to long videos: Introducing Mirasol3B, a multimodal autoregressive model for learning across audio, video, and text modalities, capable of handling longer videos and outperforming state-of-the-art approaches on video question answering and audio-video-text benchmarks.

Flag harmful content using Amazon Comprehend toxicity detection in user- or machine-generated text.

Examine 7 key themes around potential harms to mitigate imminent safety threats from AI.

Ghostbuster is a state-of-the-art method for detecting AI-generated text, achieving high performance across different domains, prompts, and models.

Google researchers discover 'Reptar,’ a new CPU vulnerability impacting Intel CPUs with potential for information disclosure and privilege escalation.